Early Serum IL-2 Levels Are Associated With CAR T-Cell Expansion Post–CAR T-Cell Therapy

Chad Williams, PhD

The early induction of cytokines within the first days of CAR T-cell therapy infusion elevated levels of serum interleukin-2 (IL-2) on days 1 and 3 and is predictive of subsequent rapid and robust CAR T-cell expansion in patients with leukemia or lymphoma treated with axicabtagene ciloleucel (axi-cel; Yescarta) or brexucabtagene autoleucel (brexu-cel; Tecartus), according to findings from a multi-study analysis that were presented at the 2024 SITC Annual Meeting.¹

Other best predictors of expansion included day 1 and 3 levels of interferon gamma (IFN-γ) and GM-CSF. Linear modeling within each individual study revealed that early IL-2 was the best predictor of CAR T-cell expansion compared with IFN-γ and granulocyte-macrophage colony-stimulating factor (GM-CSF). Furthermore, IL-2 was the only cytokine that still exhibited a strong connection to expansion at day 7.

“Serum IL-2 monitoring could allow for both early detection of CAR T-cell expansion and acute clinical activity,” first study author Chad Williams, PhD, senior scientist at Kite Pharma, and coauthors, wrote in a poster of the data.

Study Rationale

CAR T-cell expansion, particularly of CD4-positive T cells, correlates with treatment efficacy and toxicity.² However, detecting CAR T cells early following infusion is challenging, requiring sufficient immune cell counts in the peripheral blood vis PCR or flow cytometry.¹

“Establishing a reliable biomarker for early detection of CAR T-cell expansion would improve the ability to monitor real-time CAR activity in patients,” study authors wrote. “Using serum cytokines or product phenotype levies the knowledge of the product attributes or non–cell-based measurements post–CAR T-cell infusion to predict CAR expansion before it occurs in the patient.”

Study Design

This study included data from over 700 patients treated with axi-cel or brexu-cel enrolled across 7 clinical trials: cohorts 1 and 2 of the phase 1/2 ZUMA-1 trial (NCT02348216; relapsed/refractory large B-cell lymphoma [LBCL]), cohort 3 of ZUMA-1 (relapsed/refractory LBCL), cohort 4 of ZUMA-1 (relapsed/refractory LBCL), cohort 6 of ZUMA-1 (relapsed/refractory LBCL), the phase 2 ZUMA-2 trial (NCT02601313; relapsed/refractory mantle cell lymphoma), the phase 1/2 ZUMA-3 trial (NCT02614066; relapsed/refractory acute lymphoblastic leukemia), the phase 2 ZUMA-5 trial (NCT03105336; relapsed/refractory follicular lymphoma and relapsed/refractory marginal zone lymphoma), the phase 3 ZUMA-7 trial (NCT03391466; second-line LBCL), and the phase 2 ZUMA-12 trial (NCT03761056; first-line high-grade B-cell lymphoma). Investigators assessed serum levels of 27 cytokines within the first week of CAR T-cell therapy using quantitative protein analyses, then performed linear regression modeling to find cytokines within that first week that were associated with CAR T-cell expansion at peak and day 7. Linear modeling of product phenotypes assessed via flow cytometry, which included 22 T-cell subpopulations, was used to predict CAR T-cell expansion.

Linear modeling between CAR T-cell phenotype and early cytokines was also performed to determine whether certain phenotypes were associated with early, expansion-predicting cytokines. Furthermore, investigators used Least Absolute Shrinkage and Selection Operator (LASSO) modeling to validate the best biomarkers within the CAR T-cell phenotypes and early cytokines that predicted CAR T-cell expansion.

Additional Study Findings

A cross-study linear model showed that CAR T-cell product phenotypes associated with expansion are less predictive of expansion than serum cytokines. Although less-differentiated CCR7-positive and CCR7+CD45RA-positive T cells were positively associated with CAR T-cell expansion, these associations were weaker than serum cytokine levels following CAR T-cell infusion.

Furthermore, investigators found that less differentiated CCR7-positive or naive CD4-positive T cells were best associated with high levels of IL-2 at days 1 and 3. Product phenotype was more strongly associated with early levels of in vivo cytokines than CAR T-cell expansion.

Findings from LASSO modeling of day 7 CAR T-cell expansion using product phenotype and serum cytokines validated the cross-study linear modeling results, showing that early levels of IL-2 and IFN-γ were strongly associated with CAR T-cell expansion. Furthermore, although CAR T-cell product features, such as the number of CCR7-positive T cells in the CAR product, were also consistently associated with CAR T-cell expansion, these expansion levels were lower than those observed in patients with early cytokine levels.

“Factors that mediate robust CAR expansion are both numerous and complex, which could explain why biomarkers more proximal to the expansion event, like early IL-2, are better predictors of CAR expansion than product phenotype,” the authors added.

References

1. Williams C, Budka J, Zhang W, Vardhanabhuti S, Mattie M, Bedognetti D. serum interleukin-2 levels in the first 24-72 hours post CD19 CAR T cell infusion is a robust biomarker of CAR T cell expansion in leukemia and lymphoma patients treated with axi-cel or brexu-cel. Presented at: 2024 SITC Annual Meeting; November 6-10, 2024; Houston, Texas. Abstract 197.
2. Baur K, Buser A, Jeker LT, et al. CD4+ CAR T-cell expansion is associated with response and therapy related toxicities in patients with B-cell lymphomas. Bone Marrow Transplant. 2023;58(9):1048-1050. doi:10.1038/s41409-023-02016-1